If a Tree Falls in a Forest, Does It Generate an Adequate Return?

• By Daniel Rohr, CFA
• On 7:00 am EDT, Wednesday September 30, 2009

This is the second of two articles on timberland investments. In Part I, we provided an introduction to the asset class, its risk and return characteristics, and recent investment trends. In Part II, we take a look at timber supply and demand with the objective of assessing what the future may hold for timberland investments.

Supply
It's often said that before the arrival of Europeans to the New World a squirrel could jump tree to tree from the Atlantic to the Mississippi without ever touching the ground. True or not, the continent certainly had a wealth of forest resources before westward expansion kicked off in earnest. It's estimated that, in 1600, forests covered 1 billion acres or a little under half of the U.S. landmass, including Alaska.

And while population growth and industrial agriculture have undeniably transformed the country's topography over the past several centuries, the United States remains endowed with an ample supply of forest to this day. According to the latest U.S. Forest Service estimate, the U.S. now has 751 million acres of forest covering about one third of the nation's territory.

Of this sum, 514 million acres are categorized as timberland, meaning forests that are at least theoretically available as a source of wood supply. This definition excludes two categories of forest that are legally or economically severed from the wood supply: (1) reserved forestland like Adirondack Park Preserve in upstate New York or Yellowstone National Park in Wyoming and (2) unproductive forestland in environmentally marginal regions like Alaska or Arizona.

Thanks to the silvicultural efforts of foresters, the wood "inventory" on U.S. timberlands has made impressive gains over the past several decades. The most recent estimate puts the inventory at 932 billion cubic feet, up 19% from two decades ago. The density of these inventories (i.e., biomass per acre), varies considerably from region to region, reflecting differences in species, soil, and climate. Inventories are greatest (about 3,500 cubic feet per acre) in the mighty softwood forests that hug the Pacific from Washington to Northern California. Move several hundred miles inland to the Rocky Mountain region and inventories per acre fall by nearly half to around 2,000 cubic feet.

Annual growth rates also differ by region. Again, the Pacific Coast region leads the way at 69 cubic feet per acre, followed closely by the South (home to the vast majority of U.S. commercial tree plantations) at 65 cubic feet per acre. Not surprisingly, these two regions are the most important sources of the U.S. timber supply, particularly for housing end markets where together they account for roughly three fourths of the U.S. supply (South: 50%; Pacific Coast: 25%). Annual growth rates are much lower in the colder climes of the Northern and Rocky Mountain regions at 40 and 25 cubic feet per acre, respectively. These regions are consequently less important sources of supply for housing end markets, accounting for a combined 25% of supply (the Northern hardwood forests are comparably more important for paper end markets). Adding it all up, U.S. timberland inventories are growing by roughly 26.7 billion cubic feet each year, up from a growth rate of 21.9 billion cubic feet per year two decades ago.

Before moving on to discuss the demand side in greater detail, we'd be remiss not to address the significant role of Canada on the supply side (owing to its proportionately smaller population, Canada is a less important consideration on the demand side). Canada has 766 million acres of nonreserved forestland with estimated inventories of 971 billion cubic feet. But because much of this land is functionally inaccessible, estimates of total forestland available for wood supply are closer to 350 million acres, on which sit total inventories roughly half those of the U.S. As measured by Canadian government sources, the potential annual harvest on these lands (a rough proxy for growth rates) is about 8.5 billion cubic feet, which puts annual Canadian inventory growth at about one third the size of U.S. inventory growth. British Columbia, benefiting from the same ocean currents that make U.S. Pacific Northwest timberlands so productive, is the most important Canadian source of supply (about 45% of the Canadian total), followed by Quebec and Ontario (roughly 20% and 10%, respectively).

Demand
Home construction and remodeling activity consumes the largest share of the annual U.S. timber harvest--about two thirds in normal years. Since more than 90% of the homes built in the U.S. use wood-framed walls and roofs, this is far from surprising. New homes are voracious wood consumers: by one estimate, the average 2,500 square foot home contains approximately 15,800 board feet of lumber and up to 15,200 square feet of panel products. As home sizes have grown, so has housing-related demand for timber, albeit not on a one-for-one basis, since a home twice the size doesn't use twice the lumber. Lumber destined for housing consumes the highest-value timber in the forest: logs of greater diameter and without defects that could undermine the structural integrity of the end product. This category of timber is known as "sawlogs." Due to their relatively high value, sawlogs can travel significant distances economically, but the bulk find their way to regional lumber mills.

Paper and related products (corrugated packaging, tissue, and so on) are the other major source of timber demand, typically consuming about one third of the annual harvest. Paper consumes the lower value timber in the forest: logs of lesser diameter and those with substantial defects. This category of timber is known as "pulpwood." With a lower value-to-weight ratio than sawlogs, pulpwood can't travel as far economically. Consequently, timberland owners with significant pulpwood inventories are highly dependent on the financial health of local paper mills. This is a critical issue particularly in the South, which accounts for about 75% of the U.S. supply of pulpwood.

To meet the various sources of timber demand, the U.S. harvests about 16 billion cubic feet per year (assuming a healthy housing market). Set against annual supply growth of about 26.7 billion cubic feet, this would seem to suggest that the U.S. has timber supply to spare. Indeed, while inventories, as mentioned above, have been expanding at an increasing rate over the past few decades, harvest levels have barely budged. In 2006, the total U.S. harvest was 15.5 billion cubic feet--a quantity little changed from the 16.0 billion cubic feet harvest in 1996 and 15.9 billion cubic feet harvest in 1986. The growth-to-harvest ratio paints a clear picture of the apparent growth in "spare" supply. This measure has risen steadily from 1.18 in the 70s, to 1.37 in the 80s, to 1.41 in the 90s. to 1.72 by the middle of this decade, just before the full brunt of the housing downturn hammered the harvest side of the equation.

Canada too, has had theoretical "spare" supply over the past couple decades, albeit less than the U.S. Canadian harvest levels have averaged around 6.5 billion cubic feet over the two decades leading up to the housing bust, compared to a potential annual harvest of around 8.5 billion cubic feet.

Supply and Demand Outlook
From the early 1960s to the mid-1990s, the annual U.S. industrial log harvest (i.e., excluding fuelwood) increased by 3.8 billion cubic feet (roughly 40%) from 10.0 billion cubic feet to 13.8 billion cubic feet. Interestingly, the primary driver of the demand growth was not, as one might expect, housing, which contributed 1.3 billion cubic feet to the increase. Rather, it was surging paper demand, which, aided by rising newspaper and magazine circulation as well as the advent of cheap and fast printer/copier technology, added 2.5 billion cubic feet to annual demand (65% of the total increase).

From the mid-1990s, however, the annual U.S. industrial log harvest declined nearly 9% through 2006. Not surprisingly, falling paper demand seems to be the primary culprit, accounting for nearly 100% of the decline from 1996 to 2006. Looking ahead, the secular decline in paper demand seems certain to exert a persistent negative impact on the total log harvest for the foreseeable future. Moreover, we do not expect to see a material cyclical rebound in paper demand once the economy recovers: the overwhelming majority of demand lost in the current downturn will remain lost.

We expect investors in Southern and Northern timberlands will suffer most from this secular trend, since paper demand accounts for a relatively larger share of these regions' annual harvest (42% and 44%, respectively). By contrast, we expect investors in Pacific Coast timberlands will suffer less, since paper demand consumes a mere 7% to the region's annual harvest. Importantly, we do not expect the revival in housing starts to their long-run sustainable levels (which we estimate at 1.72 million) will be sufficient to offset the decline in paper demand. In other words, we expect average annual log demand in the next decade will fall below levels that prevailed in both the current and previous decades.

On the supply side, even if inventory growth rates level off (not likely given continued advancements in silvicultural practices and tree genetics, as well as the pent-up supply created by the housing slump), net annual growth is likely to exceed harvest levels for the foreseeable future, possibly pushing the growth/harvest ratio to record levels.

All told, our supply and demand forecast implies an unfavorable situation for log prices at the national level, and, consequently, timberland values. There are, however, certain regional exceptions that could brighten an otherwise cloudy outlook. In particular, investors in Pacific Coast timberlands should benefit from the supply constraints associated with the devastation wrought by the Mountain Pine Beetle in British Columbia. Add that to the fact that Pacific Coast timberlands are less dependent on paper demand, and we think the cash flow prospects for timberland investments in the region are fairly robust (capital appreciation is another matter).

In the South, which faces an otherwise gloomy outlook due to falling paper demand, the use of wood for energy offers a bit of hope. Industry research firm RISI expects to see 27 million green tons of incremental annual wood biomass demand in the South by 2015. Assuming roughly 40% of this sum would be obtained from logs (the remaining portion would come from lumber mill residues and tree trimmings), the South might see an additional 300 million cubic feet of log demand, equivalent to about 10% of the Southern pulpwood harvest in 2006. Unfortunately, given the roughly 20% decline in paper consumption we've already seen since 2006 (the exact sum varies by grade--newsprint, for example, is much worse), we do not expect rising wood biomass demand will fully compensate for lost paper demand.

Timberland Investment Returns: The Party's Over
Now that we've reviewed the supply and demand picture, let's return to one of the questions we put forth in Part I of our series on timberland investments:

With the transition from industrial to investor ownership of timberlands now nearly complete, is the capital appreciation of the past two decades a one-time, fund-flow-driven occurrence or do timberland values have more to run?

Judging by the subtitle we've assigned this section, it should come as little surprise that we think the asset class's stellar run has come to an end. To understand why this might be so requires that we examine why the ownership transition may have caused timberland values to appreciate so much in the first place.

We see three main drivers. First, any time large sums of money chase illiquid and finite assets--as has certainly been the case with timberland--asset price inflation typically ensues. Second, institutional investors assign greater value to the diversification benefits that timberland affords (i.e., vis-à-vis other financial instruments) than an industrial timberland owner. This factor alone might make timberland more valuable to an investor than a lumber mill. Third, and perhaps most important, when buyers of a given asset have a substantially lower discount rate than the sellers, prices for that asset will naturally rise. In this case, timberland buyers have had longer investment horizons than the sellers, which they argue permits them to assume greater liquidity risk than the institutional sellers. As noted in the Yale Endowment's Annual Report, "The Endowment's long time horizon is well suited to exploiting illiquid, less efficient markets such as venture capital, leveraged buyouts, oil and gas, timber, and real estate."

To illustrate this final point, let's consider a simplified example of how a seller (Company A) and buyer (Endowment X) might value a tract of timberland that generates cash flows of $50 per acre annually into perpetuity. Let's assume Company A, a large paper manufacturer, faces a cost of capital of 9%. The endowment assigns a discount rate of 6%, substantially lower than the industrial seller, but reasonable given its very long time horizon. Capitalizing the $50 per acre cash flow at these two discount rates, we see the timberland tract is worth $555 per acre to the seller and $833 per acre to the buyer. After a bit of haggling, they might agree on a price of $600 per acre: both are happy.

The following year, a very similar tract of land comes up for sale. The seller (Company B) and buyer (Endowment Y), noting the $600 per acre price of the previous transaction, may agree on a $700 per acre price. Given the disparate values the two entities assign the tract, both are happy with the price: Company B sold for $145 above its valuation and Endowment Y bought for $133 below its valuation. Endowment X is also happy, since it can make a reasonable case for marking up the value of its timberland to $700 per acre. Endowment X's 1-year return is $50 of income plus $100 of capital appreciation: roughly 25% on its initial investment of $600--not bad.

Then along come Company C and Endowment Z, making a similar deal, on a similar tract of land, at a still higher price. Endowments X and Y happily mark up the value of their holdings. The process repeats over and over until the discount rate arbitrage comes to an end.

That's just about where we find ourselves today: the massive capital appreciation that has defined the timberland return series over the past 20 years has drawn to a close. Going forward, investors will need to depend on income, not capital appreciation, for their investment returns. As implied by our supply and demand outlook, we see little reason to believe inflation-adjusted timberland income will exceed levels seen in the past two decades. In some regions--the South and North in particular--we expect falling paper demand to push real incomes lower.

That said, we are by no means suggesting timberland has no role to play in institutional portfolios. In our view, owing to the unique characteristics addressed in Part I of our timberland series, it certainly does. Rather, we would simply argue that in making asset allocation choices, decision-makers are ill-served by relying on timberland's historical returns, since the impact of ownership turnover distorts any attempt to perform risk/return analysis on the historical data set. More so than usual, the disclaimer "past returns are not indicative of future results" holds true here!

The bigger worry is that the capital appreciation tail wind might become a head wind if institutional investors sour on timberland and fund flows reverse. Indeed, pension funds and endowments seem increasingly concerned that they have misjudged the value of liquidity, which may prompt reconsideration of the discount rates applied to private equity, real estate, timberland, and other illiquid investments. As the chairman of the University of Chicago board of trustees recently told The Wall Street Journal, "We had underestimated the value of liquidity and overestimated our degree of diversification." The CIO of the Washington State Investment Board expressed similar concerns about liquidity to Bloomberg, remarking "I work for over 400,000 employees, and they can't eat IRRs. At the end of the day, I care about how much do I give you and how much money do I get back."

Sources: U.S. Forest Service, Canadian Council of Forest Ministers, RISI, American Forest and Paper Association, Wall Street Journal, Bloomberg, Company filings, Morningstar estimates

Buzz up!